Anderson’s behavioral model and the theory of planned behavior (TPB) are widely used to identify the factors associated with older adults’ willingness and behaviors to use care services. The TPB indicates that willingness is the direct predictor of individual behavior decisions, which are influenced by attitude, subjective norms, and perceived behavioral control. As a mature analytical framework with a convincing interpretation, Anderson’s behavioral model suggests that the use of medical care services is affected by needs, predisposing, and enabling factors. In this study, the TPB and Anderson’s behavioral model were integrated to form our conceptual framework (see Figure 1).

Specifically, with regard to the three sets of influencing factors in Anderson’s behavioral model, the central role of the care needs in driving CCS utilization has been evidenced in previous studies (38). Furthermore, the predisposing factors, like demographic characteristics and individual beliefs, can affect people’s propensity to use social care (18). Among them, an individual’s belief and the attitude of the TPB were merged as they have essential consistency, which reflects one’s positive or negative belief or attitude toward some behaviors and greatly determines their actual behaviors (19). In addition to the individual-level attitude, the TPB indicates that social norms, reflecting the attitude or belief from the environmental level, such as family, neighborhood, and society, to CCS may also affect one’s willingness and behaviors. Hence, we integrated social norms into the predisposing factors. Enabling factors reflect the capability of individuals or families to obtain care service resources, including personal factors such as information acquisition ability, financial factors such as incomes, and community factors such as the accessibility of services in each community. Similar to the perceived behavioral control of the TPB, which indicates the facilitative or suppressive factors perceived by people in acting on a behavior (39), enabling factors could form individuals’ positive perception and expectation of CCS utilization and further facilitate their behaviors.

All influencing factors are expected to predict the willingness of older adults to use CCSs and further impact the use behavior. For the transform status between positive willingness and actual behavior, three possible results are expected, including willingness-behavior consistency, willingness-behavior deviation (both willingness-no behavior and no willingness-behavior), and no willingness-no behavior.

A cross-sectional survey was conducted by our research team from January to April 2023 in China. Four provinces, namely Jiangsu, Hebei, Hubei, and Shaanxi, were included in this survey. These provinces cover the eastern, western, and middle regions of China. In terms of CCS development, the Chinese government has started promoting the establishment of CCS facilities since 2010 with policy and financial support. During the period of the 13th Five-Year Plan (2016–2020), the national CCSs pilot districts selected from Jiangsu, Hubei, Shaanxi, and Hebei were 10, 9, 8, and 4, ranking at the upper-middle level among 30 provinces (40). By the end of 2023, the facility coverage rate of CCSs in all four provinces has reached 90%, making it reasonable to choose them as an epitome of the general development status of CCSs in China.

The survey was fully conducted using the stratified sampling method. In each province, according to the administrative division and economic development level, three cities were selected, and the isometric random sampling method was used in the selection of two or three counties (districts) in each city. Three to four typical urban communities equipped with integrated service centers of CCSs were selected according to their administrative divisions. In each community, about 30 participants were randomly selected with the help of grassroots community workers. Specifically, the older adults were invited to participate in our survey if they (1) were aged ≥60 years, (2) were able to communicate easily or communicate with the investigators’ assistance, and (3) volunteered to participate in the study. Data collection took approximately 30 min for each participant through face-to-face interviews and surveys. A total of 1,233 older adults were included in the final sample. Informed consent was obtained from all participants prior to the interviews. Study protocols and consent forms were approved by the medical ethics committee of Health Science Center of Xi’an Jiaotong University (approval number 2016–416). All participants provided informed consent to participate in the study.

In this study, our dependent variables are the willingness and the willingness-behavior transformation of the older adults to use CCSs, which describe the matching status of the older adults’ willingness and behavior. Guided by the policies and previous literature, three CCS categories, including daily care services (DCS), health care services (HCS), and recreational and cultural services (RCS), were analyzed in this study. In terms of measurement, the willingness of the older adults to use each CCS category was measured by the question “Did you have the willingness or demand for using DCS/HCS/RCS in the past year?,” with 1 for “Yes” and 0 for “No.” Also, participants were asked if they used DCS/HCS/RCS in the past year to measure the behavior of the older adults to use each CCS category. Consequently, if the participant reported that he/she had the willingness or demand for one CCS category and had used it before, that is, he/she had both positive willingness and actual behavior, the consequence of willingness-behavior transformation was identified as “willingness-behavior consistency,” with a value of 2. If the participant reported that he/she had the willingness or demand for one CCS category but had never used it before, that is, he/she had only positive willingness, the consequence was identified as “willingness-behavior deviation,” with a value of 1. If the participant reported that he/she had no willingness or demand for one CCS category and had never used it before, the consequence was identified as “no willingness-no behavior,” with a value of 0. Another possible matching status was “no willingness-behavior,” in which a participant reported no willingness or demand for a particular CCS category, but had previously used it. However, since no participants reported this status in the survey, it was not included in the discussion of this study.

Following the conceptual framework proposed above, three sets of potential influencing factors were identified. Predisposing factors include demographic and socioeconomic characteristics, traditional older adults care beliefs, and perceived social norms. Demographic characteristics were measured by age, sex, marital status, and education level. Traditional old-age care belief (TOCB) may affect an individual’s attitude to use CCSs, which was measured by the question “Do you agree with the concept of raising children for old age,” with 1 for “Yes” and 0 for “No.” Perceived social norms include three variables of “family’s attitude,” “neighbor’s attitude,” and “policy advocacy.” Family’s attitude was measured by the item “My family members support me to use CCSs,” with 1 for “Support” and 0 for “Non-support.” Neighbors’ attitude and policy advocacy were measured by items “Older people around me often use CCSs” and “I often hear the policy advocacy about CCSs,” rated on a five-point Likert scale from very disagree = 1 to very agree = 5.

Enabling factors include an individual’s information acquisition ability (IAA), space accessibility, service affordability, and province, reflecting the accessibility of CCSs to participants. IAA and space accessibility were measured by items “I can use the Internet to search information of CCSs” and “It is convenient to get to the CCSs center from my home,” with 1 for “Yes” and 0 for “No.” Service affordability was measured by the item “The price of DCS/HCS/RCS is reasonable,” with responses ranging from disagree = 1 to very agree = 5. Given the influence of the development levels of CCSs and the economy on the social care resources, the province was measured as a regional variable in this study.

Health status and family care status were selected as needs factors. Health status was measured by the question “How do you rate your health status?” which is rated on a three-point Likert scale with bad = 1, fair = 2, and good = 3. Family care status was reflected by asking participants if they were an empty nest, with 1 for “Yes” and 0 for “No.” The detailed coding and descriptive statistics of each set of variables are shown in Table 1.

In this study, descriptive statistics analysis and regression analysis are used. First, simple descriptive statistics are estimated to illustrate the basic characteristics of participants and their willingness and willingness-behavior transformation to use CCSs. Second, for the dependent variable of “willingness,” a binary logistic model was used to investigate the influencing factors of the older adults’ willingness to use CCSs. The model estimated by the Equation 1 is as follows:

where Y denotes the probability of the older adult’s willingness to use CCSs. p represents the probability of the occurrence of willingness. β is the estimated coefficient. α is a constant term. X signifies the factor influencing the older adult’s willingness to use CCSs.

Thirdly, as the dependent variable of “willingness-behavior transformation” is a probability event, with three probable states of “willingness-behavior consistency (Y = 2),” “willingness-behavior deviation (Y = 1),” and “no willingness-no behavior (Y = 0),” the multinomial logistic regression models were used to further explore the factors influencing the willingness-behavior transformation of older adults to use CCSs. The model constructed in the Equation 2 is as follows:

where P denotes the probability of the older adult’s willingness-behavior transformation and P 1 + P 2 + P 3 = 1 . β is the coefficient of the independent variable. α is a constant term. X is the set of influencing factors.

As shown in Table 1, a total of 1,233 older adults participated in the final survey. The demographic characteristics of participants are reported as follows: among them, 42.58% held the traditional belief in raising children for old age, and 44.53% received support from family members to use CCSs. The mean scores of neighbors’ attitude and policy advocacy were 3.31 and 2.91, respectively. Additionally, 81.75% of participants reported being able to use the Internet to search for CCS information. A total of 78.59% indicated that it is convenient to access the CCS center, and the mean score for service affordability was 1.95. The mean value of self-reported health status was 1.86, and 68.61% of participants were classified as empty-nest older adults.

Table 2 shows the proportion of willingness and the three willingness-behavior transformation statuses of older adults to use CCSs. As a whole, participants showed high willingness, high willingness-behavior consistency, and high willingness-behavior deviation regarding using CCSs. 68.69% of participants showed positive willingness or demand for one or more CCSs. Among them, 33.98% of participants converted their willingness into actual behavior, while 34.71% of them had only willingness and failed to convert it into behavior.

In terms of the specific CCS category, the willingness of older adults to use CCSs showed a significant decrease. Participants having a willingness to use DCS, HCS, and RCS were 48.26, 47.85, and 36.25%, respectively. This illustrates that participants might show different willingness-behavior transformation status when using different CCS categories. Specifically, participants showing willingness - behavior deviation of using DCS, HCS, and RCS were 29.52, 36.33 and 18.73%, while participants showing willingness - behavior consistency were only 18.74, 11.52 and 17.52%, respectively, indicating the large gaps between high willingness and low behavior of older adults to use CCSs.

The binary logistics regression estimation was conducted to identify factors affecting the willingness of older adults to use DCS, HCS, and RCS. As shown in Table 3, except for age, all predisposing factors, enabling factors, and need factors had significant, albeit selective, effects on willingness to use CCS categories.

Specifically, among the predisposing factors, female older adults were more willing to use DCS compared to their male counterparts. Participants who were single showed a higher willingness to use DCS than those who were married. Older adults who were educated in high school and higher had a significantly higher willingness to use DCS and RCS. Individuals agreeing with the traditional old-age care belief were more likely to depend on family care in daily life, thus significantly restraining their demands for DCS supported by the community. As for the social norms factors, family support had a positive effect on willingness to use DCS, and both the peer effect from the neighborhood and the guidance of policies had a significantly positive relationship with older adults’ willingness to use DCS, HCS, and RCS. For enabling factors, IAA had a positive and significant effect on willingness to use HCS, space accessibility was positively related to willingness to use DCS and HCS, and service affordability was positively associated with willingness to use DCS and RCS. Compared to participants living in Hebei, participants living in Jiangsu were more likely to use three categories of CCSs, while participants living in Shaanxi showed lower willingness to use HCS. Regarding needs factors, older adults with better health status showed less willingness to use DCS and HCS, and people who were empty nesters were more willing to use DCS and RCS.

Based on these findings, effective policy implications to improve the willingness and bridge the gap between willingness and behavior can be put forward. First, the exact demand evaluation is key to the effective supply of CCSs and transforming the older adults’ willingness to use CCSs into behavior. During the evaluation, more attention should be paid to the differential characteristics of older adults to better match their service demands. For instance, the age structure of the community should be taken into consideration in the allocation of DCS resources, with more resources allocated to the advanced aging communities. Furthermore, people-targeted services such as bath assistance and housework support need to be designed for older adults who are female, single, or in an empty nest. High-quality medical services and diverse recreation service items should be improved to meet the high demands of the educated older adults.

Second, in addition to guiding the older adults’ beliefs and evoking their potential demands for CCSs, more efforts are needed to create a positive and supportive environment, especially the neighborhood and policy environment, for them to use CCSs. For instance, it is important to strengthen the promotion of CCSs’ merits in improving quality of life in later years and alleviating the burden of family care. Policy guidance should also be directed toward community-dwelling older adults with care needs, as well as their family members. Additionally, as neighborhood welfare participation plays an important role in an individual’s welfare participation (52), it is suggested to give full play to the positive neighborhood effect and peer effect of using CCSs. The key is to take improving service ability and users’ satisfaction as the core and further establish community user mutual assistance groups to take these users as important propagandists to mobilize more people.

Third, it is necessary to take measures to improve the accessibility of CCSs for community-dwelling older adults. There has been a wide consensus that Internet technology has great application prospects in widening public resource access (18). Therefore, extending the Chinese government’s ‘Internet Plus’ strategy into the field of CCSs is urgently needed, especially through the application of Internet technology in the demand collection and service delivery. This would help improve the matching of supply and demand and eliminate barriers between services and older adults. Furthermore, the strategy of the “15-min older people care service circle” should be further implemented by widening the service coverage and shortening the service distance. Additionally, a differentiated pricing strategy can be carried out with multi-grade service, and the government-paid service or welfare voucher should cover the vulnerable older adult groups. Finally, more emphasis should be placed on the high-quality development of CCSs to reinforce the competitiveness of these services, especially for healthcare services. For the government, incentive policies should be implemented to attract professional service institutions as service providers. Additionally, they can conduct regular assessments and set specialized fiscal appropriations to support and encourage the high-quality service providers.